|Remarks by Governor Laurence H. Meyer
At the Ohio Bankersí Day Conference, Columbus, Ohio
March 15, 2001
Capital Standards and Community Banks
I am pleased to be here today to discuss capital standards at community banks. This topic does not reflect any apparent problem with the current capital standard for such organizations. The system has worked well for both supervisors and community banks. With some minor exceptions, it continues to work well. Nonetheless, the issue calls out for discussion because of the new Basel Capital Accord proposed in mid-January as well as the banking agencies' advance notice of proposed rule-making last October on possible revisions to community bank capital rules.
By way of background, you may recall that the original Basel Accord, reached in 1988, was designed for large, internationally active banks. Nonetheless, regulators in the United States applied it universally to all domestic commercial banks and bank holding companies. Other countries took a similar approach.
Technology and banking practices have advanced since 1988, and especially for the large, complex banking organizations, the rule has developed serious shortcomings. Risk-measurement techniques have improved; the costs of generating, maintaining, and analyzing data have declined; and capital markets have become more tightly intertwined. As a result, the current accord that worked well before these developments has proven to be far too crude for practices today, particularly practices at large, complex banking organizations.
Supervisors in industrial nations have concluded that they need a capital measure for these large banks that is both more risk-sensitive and more compatible with current market practice than the current accord. My focus today is on what all of this means for the overwhelming majority of U.S. banks, specifically community banks.
The new Basel proposal clearly applies to all of the institutions we supervise under our program dealing with large, complex banking organizations. But that program covers only about twenty domestic entities of the nearly 9,000 U.S. banks, almost all of which have assets exceeding $40 billion. The vast majority of banks in this country are neither large nor complex and seem to be unlikely candidates for a high-maintenance capital standard that requires costly and extensive risk-management systems. More than 98 percent of U.S. commercial and savings banks, for example, have assets of less than $5 billion, and 95 percent have assets of less than $1 billion.
Last fall, in anticipation of the Basel proposal, the U.S. banking agencies asked for public comment on possible alternative standards for noncomplex banks, simpler than anything in the Basel proposal. I will be discussing the various "simpler" options--both from our advance proposal last fall and from the new Basel framework proposed in January. One outcome after considering these options may be to retain the current system, or tweak it only modestly, for most noncomplex, non-internationally active banks.
Innovations in the New Proposal
The Basel proposal represents a watershed in supervisory policy, redefining the regulatory approach to bank supervision for some banks and providing others with new incentives to improve their risk-measurement procedures. Let me call your attention to some of the most important innovations in the new proposal and, at the same time, explore whether these innovations would be appropriate for community banks.
Three pillars. The Basel Committee's approach rests on three so-called pillars: (1) the new capital standard itself; (2) increased supervisory review of banks' internal assessments of their own capital adequacy; and (3) additional public disclosure of bank risk profiles. In contrast, the original Basel Accord rested on a single pillar, a regulatory minimum for capital.
Pillars 1 and 2 in effect distinguish two concepts of capital adequacy: regulatory minimum capital set by pillar 1 and the supervisor's evaluation of banks' internal calculations of their economic capital under pillar 2. By "economic" capital I mean the amount of capital that management has determined is required to reflect the risks, strategy, and objectives of its individual bank, regardless of the regulatory structure. The regulatory minimum, in effect, provides a threshold for early intervention by supervisors. The procedures and assumptions embedded in the internal measure of economic capital drive business decisions of the bank by permitting management to gauge the capital costs of, and better establish the price for, credits of varying degrees of risk. Using pillar 2, supervisors would plan to evaluate the internal capital- allocation process for integrity and economic sense, and to ensure that the underlying processes and assumptions used to measure economic capital are reasonable and consistent with the principles in pillar 1. Finally, pillar 3 looks to market discipline to supplement and reinforce the effectiveness of pillars 1 and 2 in disciplining the risk-taking of banking organizations.
In principle, the three-pillar approach applies to community banks as well as to LCBOs. In practice, however, the relative importance and detailed application of the three pillars would clearly be quite different at LCBOs and at community banks. It would be difficult and expensive to develop and maintain the risk-management infrastructure necessary to evaluate a customer's creditworthiness with the precision and internal quality controls necessary to use the more-advanced capital approaches in the Basel proposal. More to the point, virtually all of these techniques and procedures seem currently unnecessary for most U.S. banks, and they will remain so unless the economics and the nature of community banking substantially change. The scale and operations of community banks neither separate the management from credit decisions nor involve the complexity that characterizes the process at larger entities. As a result, a simpler approach for defining the regulatory minimum capital is likely to continue to work well for community banks, even as we need to move to a more complex and more risk-sensitive approach for LCBOs.
In the United States, supervisory oversight--pillar 2 under the new proposal--has long been an integral part of our regulatory framework, both for LCBOs and community banks. For both sets of institutions, supervision has evolved to a more formal, risk-focused approach, with our exams increasingly emphasizing the adequacy of internal processes and controls.
Since long before the risk-focused approach was formalized, however, examiners have been rating a bank's capital based on their assessment of the risk of the institution. For example, a bank with stated capital ratios well in excess of minimum requirements--or in excess of the "well capitalized" threshold under today's standard--may still be assigned a less-than-satisfactory rating for the capital component of the CAMELS rating if, in the opinion of the supervisor, the risk at that institution warrants such a rating. In 1991, with the passage of the Federal Deposit Insurance Corporation Improvement Act, the Congress formally recognized the role of supervisory oversight in assessing capital adequacy of banks in relation to risk, permitting federal bank regulators to reclassify a bank's capital category under certain conditions. Thus, for example, a federal bank regulator may reclassify a "well capitalized" bank to the "adequately capitalized" category if the bank is in unsafe and unsound condition or if it fails to correct unsafe and unsound practices. The supervisory oversight of capital adequacy, to be sure, is more intense and continuous in the case of large, complex institutions, but what is now being called pillar 2 has long been, and will continue to be, an integral part of the framework for community banks.
Market discipline (pillar 3), however, may be particularly problematic for community banks, since such institutions are often closely held and not rigorously and continuously evaluated by market analysts and investors. Even if they are traded on a regional basis, the depth of the trading is likely to be inadequate to provide timely signals to supervisors and discipline to management. So, though pillar 3 is not irrelevant for community banks, it is unlikely to be as effective as it is for large, complex banks.
Multilayered approach. As I have noted, the proposed Basel capital standard is highly complex, but for good reason. The proposal contains overlapping options that offer several approaches to accommodate banks with varied resources, expertise, and risk profiles. The question is whether the differentiation goes far enough for the U.S. banking structure.
The most basic approach under the proposed accord is the so-called standardized method. The standardized approach is conceptually the same as the current capital accord, but more risk-sensitive in some respects. Under all the Basel approaches, a bank, as now, would allocate a risk-weight to each of its assets and off-balance-sheet positions and calculate a sum of the risk-weighted asset values. The capital requirement remains 8 percent of the institution's risk-weighted assets. The primary change, under the standardized approach, would be to base the asset risk-weights for exposures to sovereigns, other banks, and corporations on the borrower's external credit ratings, when possible. The proposal would also give more recognition to a bank's efforts to mitigate risk through guarantees, collateral, credit derivatives, and netting agreements.
Those changes may be attractive to many institutions; however, the clear focus of the new Basel proposal is the basing of regulatory capital requirements on a bank's own internal credit ratings, the so-called internal rating based or IRB approach. There are two variants of the IRB approach: the "foundation" and "advanced" options. Both were designed for banks with the resources and expertise to build and maintain sophisticated risk-management systems and for which the current capital standard more seriously mismeasures risk.
The difference between the two IRB methods is that the foundation approach is designed for banks with systems that do not permit full reliance on internal risk estimates for calculating capital requirements. Under the foundation approach, missing risk parameters would be supplied by the supervisor and would be set conservatively. If banks believe these conservative estimates are inappropriate or too large, they can invest in the research and systems to prove the point and pursue the advanced IRB approach. Banks using the advanced approach would have more freedom, but their methodologies would first be closely reviewed and approved by the supervisor.
This multilayered structure offers a flexible approach that could apply to a wider group of banks, not just to large, internationally active banks. The key question for U.S. community banks and their regulators is whether any of the options in the new Basel proposal--especially the standardized approach--is a good match for community banks. Another question is whether competitive pressures would provide incentives--perhaps overwhelming incentives--for community banks to pursue an IRB approach. Would such an allocation of resources by community banks and supervisors be desirable and efficient? Finally, if not one of the Basel alternatives, what capital standard would be most appropriate for community banks in the United States?
Comprehensive treatment of risk. Although originally portrayed as dealing only with credit risk, the standard covers, in effect, virtually all banking risks through the experience-based, but unscientific, capital charge of 8 percent. That is, although the current standard was structured on the basis of credit risk alone, its level was set higher than warranted for just credit risk, in effect incorporating a "buffer" that could be used to cover "other risks."
As we get more precise in measuring credit risk, we lose some of the buffers previously built into the standard to cover operational and other risks. Explicitly or implicitly, we now need to make judgments regarding the respective correlations of different risks. Are they additive or independent, and to what degree are they one or the other? Can the same capital cover all of them, or do banks need capital for each one? But again the question is whether the current approach, with its less precise measurement of credit risk and its buffers for other risks, is working well enough for community banks and should be left as it is.
The Bifurcation Alternative: The Interagency Advanced Notice of Proposed Rulemaking
As I have mentioned, community banks in the United States have, besides the new Basel proposal, another capital proposal to consider. I am referring to the advance notice of proposed rulemaking issued last October by the U.S. banking and thrift agencies that explores alternatives for simplifying the capital standard for noncomplex, non-internationally active U.S. depository institutions. This is the so-called bifurcated approach. The advance proposal suggested three possible ways to structure a capital standard for non-complex banking organizations. The first option would combine a risk-based ratio with a leverage ratio. This option would retain much of the current approach while tailoring capital charges to the size, structure, and risk profile of less- complex banking organizations. The second option would use only a leverage ratio. Taking an approach not unlike the primary capital standard we used throughout the 1980s, it would drop any semblance of a risk-weighted measure and base a bank's capital requirements only on the ratio of its capital to total assets. The third alternative is simply a modified leverage ratio that would add the exposures arising from loan commitments and other off-balance-sheet transactions to total bank assets.
Reflecting its early stage of development, the proposal presents more questions than answers and does not flesh out any of the three alternatives it presents. One important question relates to how noncomplex institutions would be defined. The proposal suggests basing the decision on a bank's asset size, the nature of its activities, and its risk profile--for example, assets less than $5 billion, moderate amounts of off-balance-sheet transactions that are mostly loan commitments, and a focus on traditional banking activities. In effect, the size threshold would be the determining factor for most banks, permitting all but a few hundred of the nearly 9,000 federally insured commercial and savings banks to calculate their capital requirements using a simple approach. Almost all thrifts as well would likely qualify for the simpler approach.
This seems to be the proper time to consider a simpler approach for community banks as we undertake major revisions to the Basel Accord. However, to be consistent with the spirit of Basel, the agencies proposed that a simpler system should also be more conservative, requiring at least a slightly higher minimum capital charge. Such a higher minimum would still, however, be below the actual (or economic) capital now maintained by virtually all community banks.
The higher capital ratios now maintained at smaller banks reflect certain economic realities. These banks generally carry greater risk than larger banks because of their greater concentrations in exposures and funding sources, byproducts of their limited geographic scope and scale. In addition, because of their more limited access to capital markets, community banks have fewer options for augmenting capital when it is needed. The higher capital ratios at community banks, compensating for these greater risks, result as much from bank managers' tolerance for risk and market expectations for banks' capital as from the influence of bank supervisors. Bank management, the market, and the supervisors all understand that the current capital standard sets minimum capital requirements, not the economic levels that are required for prudent operations.
For all these reasons, community banks have consistently maintained higher capital ratios than larger banks since long before a regulatory capital standard was implemented. At the end of 2000, for example, banks with total assets of less than $1 billion had an average tier 1 leverage ratio of almost 10 percent and an average total risk-based capital ratio of almost 15 percent--both about double the minimum standard. In contrast, banks with assets greater than $10 billion had average ratios of about 7 and 11.5 percent, respectively. These and often larger differences extend back as far as data are available.
Despite having capital ratios typically well above any minimum that regulators would require, community banks appear to have little interest in the advance proposal for noncomplex banks. In the small number of comment letters we have received, community banks did not express a clear preference for gaining a simpler regulatory capital measure in exchange for potentially higher minimum capital requirements--requirements still well below their economic or maintained capital. The gains in simplicity were apparently not viewed as offsetting the costs of changing from the current set of rules to a purportedly simpler system. The intent of the proposal, by the way, was to reduce regulatory burden on smaller banks while ensuring that their capital remains at prudent levels.
I hope my comments have clarified that the new Basel and interagency proposals raise some important questions about the appropriate capital standard for U.S. community banks. If community banks want to shape the outcome, they must convey their views on several key questions.
Who is under what standard? A crucial question that we should consider is which banks or types of banks will pursue which alternative capital standards. Clearly, all or almost all of the truly global institutions, worldwide, would be expected to use an IRB approach, with most of them moving to the advanced approach, if they are not there at the outset. Most other large banking organizations, such as the large U.S. regional banks, are also likely to pursue that approach, at least eventually.
Supervisory and market pressures, alone, would presumably force the global and large regional banks in that direction. Given the systemic risk posed by implications and sheer complexity of many of our largest institutions, supervisors worldwide should at least be urging these firms in that direction.
However, if the standard is designed and calibrated properly, the incentives should include both a carrot and stick. Institutions that meet the demanding risk-management and public-disclosure standards required under the IRB approach deserve some consideration in the form of regulatory minimum capital charges more accurately reflecting the underlying risks. The IRB banks' risk-measurement and management processes will be strengthened, and capital markets will have more information with which to judge risks and arguably assert more discipline, if risks appear to rise. The key is to align the incentives--capital charges and costs--just right and to ensure, largely through the supervisory process, that the required operating standards are met.
Is the standardized approach a good fit for community banks? Perhaps the larger question relates to which alternative most community and smaller regional banks will choose. Most of them, I suspect, would opt for the standardized approach, or a still-simpler method, as suggested in the bifurcated approach proposals of last fall. Or, as noted, many might just prefer the existing rule in order to avoid transition costs. Of course, transition costs should be balanced against the resultant regulatory burden and capital charges. If, for example, the incentives for banks to move to more-advanced standards are exceptionally attractive, a larger number of community and regional banks may choose to bear these costs. My own judgment is that will not be the case for most community banks. From the supervisors' point of view, any incentives to move to the more-advanced approaches should be commensurate with the identified risks and should not materially disrupt the terms of competition.
Although the alternative was not specifically highlighted, is remaining on the current capital standard a reasonable option to consider for community banks? The standardized approach is only a modest change from the current system, and many of the other changes are not relevant to community banks. Community banks, for example, presumably do not have many externally rated credits, nor do they generally lend to foreign governments. Does it make sense then to require them to change their system to be compatible with the standardized approach, resulting in some costs, when the benefits may be minimal?
One possible solution may be to think of the current system as one of the simpler regimes for noncomplex banks--that is, simpler than the new Basel options. This would allow community banks to remain under today's system with the same regulatory capital minimum.
Would the current system need to be tweaked to minimize competitive pressures? I have heard some community bankers say that competitive pressures would force them to follow larger banks to an IRB standard. That is certainly not our intent, and I am not convinced that result is inevitable. If it is, we need to look more closely at how well the incentives are balanced.
The few respondents to the interagency advance proposal for noncomplex banks suggested some revisions to the current standard. Specifically, they suggested reduced requirements on certain assets, such as low loan-to-value mortgages and certain collateralized consumer loans. In such cases, they claimed, much less capital is needed than is required, and most supervisors would probably agree. The difficulty is that the respondents did not also suggest which exposures, in turn, should get higher capital requirements, or how high-risk exposures could be as easily and objectively identified in community banks. Simply cherry-picking a bank's best assets to receive lower requirements is not workable. Indeed, since supervisors do not believe that overall capital requirements are excessive, you should not expect to be offered lower risk-weights for low-risk assets without also accepting higher risk-weights for higher risk activities. At the very least, that might mean a higher risk-weight--possibly far higher on the basis of the historical record--for classified assets. It could also mean a higher risk-weight for subprime and generally riskier loans.
In closing, I would emphasize that the current standard appears to be working well for community banks. As with any change in regulation, an appropriate cost-benefit analysis must be undertaken. That is, do the benefits from any new standard more than outweigh the added costs of any increase in complexity as well as the one-time cost of adjusting to a new system? The proposed Basel Accord appears to be well designed to remedy the deficiencies of the current standard for large, complex banking organizations. The U.S. banking agencies, however, will have to give careful thought as to whether any of the options in the Basel proposal are a good fit for U.S. community banks. If they are not, we need to consider what simpler standard would be more appropriate. That was the rationale for the advanced notice of proposed rulemaking. Remaining on the current standard (with or without some minor modifications) is an option we will need to consider for community banks. And we need to hear your views.